Surron—Study of Evaporation from Water-Surfaces. 167 
Only ratios must be considered in this Table, for the observa- 
tions with the two pipes were not made simultaneously. It will 
be understood, of course, that the air near the ground inside the 
screen containing the pipes is colder than it is at one and two feet 
higher, 7.¢., at the levels of the mouths of the pipes, especially at 
night, and consequently that the interchange of air below may 
not be very vigorous. Moreover, the air-channels below the flanges 
are very small. In spite of this we see from Table 20 that our 
modification of the experiment has considerably altered the diurnal 
variation of the rate of evaporation in the pipes, transforming it 
into something more nearly like that of the open cup than it was 
before. We see, in fact, that the hourly rate of evaporation 
between 5 and 8 p.m., instead of being (Table 10) two and a half 
times as great as that between 8 a.m. and 5 p.m., is now only 
about half as great again. 
If the curious, dissimilar, diurnal curves of evaporation in the 
metal pipes, from the tub and from the open cup, be really due in 
great part to convection currents, the same cause may be respon- 
sible for the want of symmetry in the diurnal curve of evaporation 
from the large tank. 
In Table 21 (page 168) are given the hourly average quantities 
of water evaporated from the tank for each month, derived 
from observations made between August, 1899, and July, 1906 
—seven years. Generally speaking, in the winter there are 
three maximum points on the curve. The first is about the 
time of sunrise,’ and is sometimes very pronounced; the second 
is the normal maximum in the afternoon; the third comes some- 
where between 9 and 11 p.m. In the summer the first and third 
maxima become faint or abortive. The first minimum points are 
very curiously arranged. In the summer they group themselves 
about the hour before sunrise; but in the winter-half of the year 
they plainly work back almost to midnight. The consequence is 
that the average time of the first minimum for the year falls about 
4 a.m., 7.¢., nearly two hours earlier than the minimum of either 
air-temperature or wind-velocity, or the maximum of relative 
1 There are trees to the east of the tank, and in consequence the rays of the sun 
do not reach the water for some considerable time after sunrise, more especially in 
the winter. The average evaporation at sunrise in July is sometimes greater than the 
normal maximum in the early afternoon. 
